File: divergence-at-use.ll

package info (click to toggle)
llvm-toolchain-16 1%3A16.0.6-15~deb12u1
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 1,634,792 kB
  • sloc: cpp: 6,179,261; ansic: 1,216,205; asm: 741,319; python: 196,614; objc: 75,325; f90: 49,640; lisp: 32,396; pascal: 12,286; sh: 9,394; perl: 7,442; ml: 5,494; awk: 3,523; makefile: 2,723; javascript: 1,206; xml: 886; fortran: 581; cs: 573
file content (20 lines) | stat: -rw-r--r-- 651 bytes parent folder | download | duplicates (2) 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
 
; RUN: llc -march=amdgcn -mcpu=gfx900 -amdgpu-atomic-optimizations=true < %s | FileCheck %s
; RUN: llc -march=amdgcn -mcpu=gfx900 -amdgpu-atomic-optimizations=true < %s -use-gpu-divergence-analysis | FileCheck %s

@local = addrspace(3) global i32 undef

define amdgpu_kernel void @reducible(i32 %x) {
; CHECK-LABEL: reducible:
; CHECK-NOT: dpp
entry:
  br label %loop
loop:
  %i = phi i32 [ 0, %entry ], [ %i1, %loop ]
  %gep = getelementptr i32, ptr addrspace(3) @local, i32 %i
  %cond = icmp ult i32 %i, %x
  %i1 = add i32 %i, 1
  br i1 %cond, label %loop, label %exit
exit:
  %old = atomicrmw add ptr addrspace(3) %gep, i32 %x acq_rel
  ret void
}